A5100439576- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Robotics and Sensor-Based Localization
- Human Pose and Action Recognition
- Hand Gesture Recognition Systems
- Medical Image Segmentation Techniques
- Multimodal Machine Learning Applications
- Image and Signal Denoising Methods
- Visual Attention and Saliency Detection
- Medical Imaging Techniques and Applications
- Infrastructure Maintenance and Monitoring
- Advanced Image Processing Techniques
- Robotic Path Planning Algorithms
- Image Enhancement Techniques
- Domain Adaptation and Few-Shot Learning
- Inflammasome and immune disorders
- Magnetic Field Sensors Techniques
- Remote Sensing and LiDAR Applications
- Streptococcal Infections and Treatments
- Thermal Analysis in Power Transmission
- Gait Recognition and Analysis
- Vehicle License Plate Recognition
- Pneumocystis jirovecii pneumonia detection and treatment
- Remote-Sensing Image Classification
- Nutrition, Genetics, and Disease
Anhui University
2024-2025
Hubei University
2025
Wuhan University
2013-2024
State Intellectual Property Office
2024
China Mobile (China)
2024
State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing
2021-2024
Cornell University
2022
The Ohio State University
2021-2022
Dalian Maritime University
2022
KTH Royal Institute of Technology
2019
This article presents a transformer and convolutional neural network (CNN) hybrid deep for semantic segmentation of very high resolution (VHR) remote sensing imagery. The model follows an encoder–decoder structure. encoder module uses new universal backbone Swin to extract features achieve better long-range spatial dependencies modeling. decoder draws on some effective blocks successful strategies CNN-based models in image segmentation. In the middle framework, atrous pyramid pooling block...
Self-driving cars have experienced rapid development in the past few years, and Simultaneous Localization Mapping (SLAM) is considered to be their basic capabilities. In this article, we propose a direct vision LiDAR fusion SLAM framework that consists of three modules. Firstly, two-staged visual odometry module, which frame-to-frame tracking step, an improved sliding window based thinning proposed estimate accurate pose camera while maintaining efficiency. Secondly, every time keyframe...
Vehicle Re-IDentification (Re-ID) aims to retrieve the most similar images with a given query vehicle image from set of captured by non-overlapping cameras, and plays crucial role in intelligent transportation systems has made impressive advancements recent years. In real-world scenarios, we can often acquire text descriptions target through witness accounts, then manually search queries for Re-ID, which is time-consuming labor-intensive. To solve this problem, paper introduces new...
Low-Light Image Enhancement (LLIE) task tends to restore the details and visual information from corrupted low-light images. Most existing methods learn mapping function between low/normal-light images by Deep Neural Networks (DNNs) on sRGB HSV color space. Nevertheless, enhancement involves amplifying image signals, applying these spaces with a low signal-to-noise ratio can introduce sensitivity instability into process. Consequently, this results in presence of artifacts brightness...
Patients with diabetic foot ulcers (DFU) suffering from severe lower limb ischemia face the risk of amputation. Concomitant oxidative stress and hyperinflammation commonly manifest within tissue affected by DFU, exacerbating deterioration DFU wounds. One-two punch strategy anti-oxidative damage plus anti-inflammatory is anticipated to tackle challenge non-healing Here, we introduced a dual-approach treatment involving probiotic Weissella cibaria (WC) modified desferrioxamine (DFO). This...
To observe the effect of acupuncture on gastrointestinal motility, synaptic structure, brain-gut peptide and 5-hydroxytryptamine (5-HT) system in rats with functional dyspepsia (FD) depression, so as to explore its mechanism underlying improvement FD depression-like behavior. Forty male SD were randomly divided into control, model, medication (fluoxetine) groups 10 each group. The model depression comorbidity was established by multi-factor stress conditioning (chronic unpredictability...
In this work, we propose a new deep convolution neural network (DCNN) architecture for semantic segmentation of aerial imagery. Taking advantage recent research, use split-attention networks (ResNeSt) as the backbone high-quality feature expression. Additionally, disentangled nonlocal (DNL) block is integrated into our pipeline to express inter-pixel long-distance dependence and highlight edge pixels simultaneously. Moreover, depth-wise separable atrous spatial pyramid pooling (ASPP) modules...
Vanilla models for object detection and instance segmentation suffer from the heavy bias toward detecting frequent objects in long-tailed setting. Existing methods address this issue mostly during training, e.g., by re-sampling or re-weighting. In paper, we investigate a largely overlooked approach -- post-processing calibration of confidence scores. We propose NorCal, Normalized Calibration segmentation, simple straightforward recipe that reweighs predicted scores each class its training...
has been demonstrated to have the strongest association with periodontitis. Within host,
What kind of emotional experience does the application digital technology in museums create for museum visitors? Can it be measured accurately and real-time? are its characteristics? This paper utilizes EEG signals PAD model as research methods to conduct real-time measurement visitors’ experiences at Tianyi Pavilion Museum Ningbo City, focusing on their physiological psychological reactions.The results show that: (1) In a quasi-experimental environment, linear SVM, polynomial kernel...
Semantic occupancy prediction aims to infer dense geometry and semantics of surroundings for an autonomous agent operate safely in the 3D environment. Existing methods are almost entirely trained on human-annotated volumetric data. Although high quality, generation such annotations is laborious costly, restricting them a few specific object categories training dataset. To address this limitation, paper proposes Open Vocabulary Occupancy (OVO), novel approach that allows semantic arbitrary...
The article proposed a method that suggest way to replace some lower category identification capacity items from the ICTCLAS segmentation result by drawing feature owns better 2-gram improve classification effect of ICTCALS method. By using KNN categorization algorithm and Naive Bayes text method, it proved this worked well on FuDan university corpus. And also analyzed reason why was relatively noneffective Sogou laboratory corpus through test.
In this paper, we present BodyTrak, an intelligent sensing technology that can estimate full body poses on a wristband. It only requires one miniature RGB camera to capture the silhouettes, which are learned by customized deep learning model 3D positions of 14 joints arms, legs, torso, and head. We conducted user study with 9 participants in each participant performed 12 daily activities such as walking, sitting, or exercising, varying scenarios (wearing different clothes, outdoors/indoors)...
In this study, we developed a deep learning model based on GD-DTPA-enhanced MRI data to predict the overall survival (OS) of patients with HCC. Our results showed that 3D-CNN can non-invasively OS The combined integrating score and clinical factors higher predictive value than models may be more useful in guiding treatment decisions improve prognosis
With the continuous increase of voltage grade in substation, switch operating overvoltage, especially very fast transient overvoltage (VFTO), and electromagnetic interference transformer substation caused by VFTO has attracted more attention. Based on a new generation intelligent 330kV demonstration project, which highest level China, this paper analyses physical process propagation high combining simulation calculation with test detection. Taking frequency model acquisition unit based Roche...